Contemplating self-sacrifice 1
In press, Journal of Personality and Social Psychology
Contemplating the ultimate sacrifice:
Identity fusion channels pro-group affect, cognition, and moral decision-making
William B. Swann, Jr. *
Angel Gómez **
Michael D. Buhrmester ***
Lucía López-Rodríguez ****
Juan Jiménez **
Alexandra Vázquez **
* University of Texas at Austin
** Universidad Nacional de Educación a Distancia
*** University of Oxford, Queen’s University, Belfast
**** Universidad de Almería
Keywords: Identity Fusion, morality, moral judgment, self-sacrifice, social identity
Author note
We thank Josh Greene and Jonathan Jong for helpful comments regarding this
article. For funds that facilitated this research, we acknowledge National Science
Foundation grant #BCS-1124382 to William B. Swann, Jr., Spanish Ministry of
Economy and Competitiveness grant #PSI2012-30921 to Angel Gómez, UNED 2013004-UNED-PROY grant to Alexandra Vázquez, and postdoctoral support from ESRC
(REF RES-060-25-0085) to Michael D. Buhrmester.
Contemplating self-sacrifice 2
Abstract
Although most people acknowledge the moral virtue in sacrificing oneself to save others,
few actually endorse self-sacrifice. Seven experiments explored the cognitive and
emotional mechanisms that underlie such endorsements. Participants responded to one of
two moral dilemmas in which they could save five members of their country only by
sacrificing themselves. Over 90% of participants acknowledged that the moral course of
action was to sacrifice oneself to save others (Experiment 1), yet only those who were
strongly fused with the group preferentially endorsed self-sacrifice (Experiment 2-7). The
presence of a concern with saving group members rather than the absence of a concern
with self-preservation motivated strongly fused participants to endorse sacrificing
themselves for the group (Experiment 3). Analyses of think aloud protocols suggested
that saving others was motivated by emotional engagement with the group among
strongly fused participants but by utilitarian concerns among weakly fused participants
(Experiment 4). Hurrying participants’ responses increased self-sacrifice among strongly
fused participants but decreased self-sacrifice among weakly fused participants
(Experiment 5). Priming the personal self increased endorsement of self-sacrifice among
strongly fused participants but further reduced endorsement of self-sacrifice among
weakly fused participants (Experiment 6). Strongly fused participants ignored utilitarian
considerations, but weakly fused persons endorsed self-sacrifice more when it would save
more people (Experiment 7). Apparently, the emotional engagement with the group
experienced by strongly fused persons overrides the desire for self-preservation and
compels them to translate their moral beliefs into self-sacrificial behavior.
Contemplating self-sacrifice 3
Contemplating the ultimate sacrifice:
Identity fusion channels pro-group affect, cognition, and moral decision-making
“I´m only really concerned about myself”
Fan Meizhong on why he callously abandoned his pupils after a
catastrophic earthquake struck his school (Spencer, 2008)
“Professor Librescu died as he lived, devoted to his students and to his profession.”
From a memorial to Dr. Liviu Librescu, who sacrificed his life to save his
students during the April 16th, 2007, Virginia Tech campus shootings.
When people encounter life and death moral dilemmas, they almost invariably
know what they should do. Nevertheless, cases like Fan Meizhong’s demonstrate that
knowing does not guarantee doing. Fan acknowledged on his blog that he knew that he
should have helped his students evacuate. Yet, when the earthquake struck, he shouted
"Stay calm, it's an earthquake!" and ran for it – without even bothering to see if his pupils
were following. Later, when one of his students asked for an explanation, he helpfully
replied, “I have a very strong sense of self-preservation”. In contrast, when Dr. Librescu
heard gunshots, he swiftly barricaded himself against the classroom door, giving his
students time to escape through an open window. This act of valor cost him his life.
Which raises a question: If merely recognizing that group members are in mortal danger
is not enough to motivate many people to sacrifice themselves to save others, what does
compel such actions?
In this article, we suggest that one important antecedent of self-sacrifice for one’s
group is identity fusion, which denotes a visceral sense of oneness with the group. We
suggest further that the sense of connection experienced by strongly fused persons
triggers emotional reactions to the plight of fellow group members, emotional reactions
Contemplating self-sacrifice 4
that motivate the decision to sacrifice their lives for these group members. Weakly fused
persons do not have the same emotional reactions when they learn that their fellow group
members are imperiled; instead their first impulse is to preserve their own lives. To test
these possibilities, in seven investigations we explored the mental processes that unfold
as people contemplate sacrificing themselves for members of their group. Identity fusion
theory (Swann, Jetten, Gómez, Whitehouse, & Bastian, 2012) provides the conceptual
context for this work.
Identity Fusion
The identity fusion approach assumes that when people become fused with a
group, they theoretically experience a visceral sense of “oneness” with that group. This
sense of oneness involves the union of the personal self (viz., idiosyncratic features of the
individual) and a social self (viz., features the individual shares with the group). In its
embrace of the distinction between the personal and social self, identity fusion resembles
“group identification”, a key variable within social identity formulations (e.g., Ellemers,
Spears, & Doosje, 1997; Postmes, Haslam, & Jans, 2013; Smith, Seger, & Mackie, 2007).
A central difference between identification and fusion, however, involves the nature of
the “glue” that bonds group members to the group. Social identity formulations hold that
group members are bound to one another through categorical ties that are based on the
degree to which members embody the prototypic qualities of the group; the unique
relationships that group members establish with one another are not regarded as an
important contributor to identification. In fact, in the widely used “minimal-group
paradigm”, participants become identified with the ingroup and biased against the
outgroup despite never having encountered a single member of either group (e.g., Billig
Contemplating self-sacrifice 5
& Tajfel, 1973; Turner, Sachdev, & Hogg, 1983). Not surprisingly, then, the pro-group
sentiments of highly identified persons are directed toward the abstract category rather
than individual members of the category (e.g., Tajfel & Turner, 1979; Turner, Hogg,
Oakes, Reicher, & Wetherell, 1987).
In contrast, the relational ties principle of fusion theory (Swann et al., 2012) holds
that strongly fused persons are not bound to the group category only, but also to their
fellow group members (see also Aron, Aron, Tudor, & Nelson, 1991; Markus &
Kitayama, 1991). The presence of such relational ties means that fused persons view
other group members as unique individuals rather than interchangeable members of the
category. These relational ties may contribute to feelings of family-like mutual obligation
and shared strength. This feelings may help explain why measures of identity fusion
consistently outperform measures of identification in predicting endorsement of fighting
and dying for individual members of the group (Gómez, Brooks, Buhrmester, Vázquez,
Jetten, & Swann, 2011a). Mutual obligation and shared strength may also explain why,
when presented with an opportunity to donate to fellow Spaniards in need of financial
help, strongly fused persons donate more personal funds than weakly fused persons
(Swann, Gómez, Huici, Morales, & Hixon, 2010b). Moreover, when strongly fused
persons are ostracized by fellow group members, they refuse to abandon the group, and
instead increase their endorsement of extreme sacrifices for the group (Gómez, Morales,
Hart, Vázquez, & Swann, 2011b). Further, in several versions of the trolley dilemma
(Foot, 1967), fused persons endorsed plunging themselves in front of a speeding train to
save several members of their group. In contrast, highly identified persons were generally
Contemplating self-sacrifice 6
unwilling to sacrifice themselves for ingroup members (Gómez et al., 2011a; Swann,
Gómez, Dovidio, Hart, & Jetten, 2010a).
Of particular relevance to our interest in the processes that give rise to endorsement of
extreme pro-group behavior, several studies showed that the effects of fusion could be
amplified either by increasing their physiological arousal (Swann et al., 2010b) or
priming people’s personal self-views (Gómez et al., 2011a; Swann, Gómez, Seyle,
Morales, & Huici, 2009). Similarly, in one study, group-directed personal agency (e.g., “I
am responsible for my group’s actions”) mediated the relationship between identity
fusion and pro-group behavior (Gómez et al., 2011a). In another, relational ties (e.g.,
“Members of my country are like my family to me”) mediated the relationship between
identity fusion and endorsement of pro-group behavior (Swann et al., in press). In all of
the foregoing research, measures of identification were relatively poor predictors of
extreme pro-group outcomes and agency did not mediate the effects of identification on
pro-group behavior. Together, this evidence suggests that identity fusion it is
qualitatively distinct from social identification (for further discussion, see Buhrmester et
al., 2012; Gómez et al., 2011a; Swann et al., 2012).
Of particular relevance here, evidence that increasing physiological arousal amplifies
the effect of fusion on pro-group behavior (Swann et al., 2010b) suggests that people’s
emotional responses may play a role in motivating fused persons to enact pro-group
behavior, including behaviors that have moral overtones. The role of emotion in moral
decision-making has recently attracted the attention of several researchers. For example,
after a decades-long emphasis on the impact of relatively controlled, explicit beliefs in
moral judgment (Kohlberg, 1969; Turiel, 1983), researchers have now provided evidence
Contemplating self-sacrifice 7
for the importance of automatic, intuitive processes in moral behavior (Blair, 1995;
Haidt, 2001; Pizarro & Salovey, 2002).
A case in point is “moral dumbfounding”, in which people blithely endorse moral
judgments that subsequent questioning reveals they have no rational basis for having
made (Haidt, 2001). Such immediate, non-conscious, intuitive processes share many of
the properties of emotional reactions. Evidence of the role of such processes in moral
reasoning has persuaded several researchers to propose that two separate mental systems
(Chaiken & Trope, 1999; Posner & Snyder, 1975) govern moral decision making, one
that is rapid, automatic, and emotion-driven and another that is deliberate and logical
(e.g., Greene, 2007; 2009; Greene, Morelli, Lowenberg, Nystrom, & Cohen, 2008;
Lombrozo, 2009). Recent evidence using Jacoby’s (1991) process dissociation model has
offered direct support for the operation of these dual systems. Whereas some moral
judgments (viz., those involving deontological principles) were uniquely related to
emotional processes such as empathic concerns and religiosity, other moral judgments
(viz., those involving utilitarian principles) were uniquely related to need for cognition.
Furthermore, just as a time-pressure manipulation selectively reduced adherence to
utilitarian principles, an empathy enhancing manipulation selectively increased adherence
to deontological principles (Conway & Gawronski, 2013).
Here we suggest that a dual process approach can also be used to illuminate the
unique moral judgments of persons who are strongly versus weakly fused to their group.
That is, insofar as fusion taps a visceral feeling of oneness with the group, threats to other
group members may activate the same emotions of tension, distress, and anxiety that they
would experience were they themselves in danger. Moreover, these emotional reactions
Contemplating self-sacrifice 8
should compel them to apply deontological principles to the plight of their group
members and resolve to save them. Among the many who decide to sacrifice themselves,
the implications of their own death should come to mind only after contemplating the
plight of the group. And if they decide not to sacrifice themselves, they should do so
while resigning themselves to find another way to save the ingroup (i.e., deny the
implications of the binary options in the dilemma).
In contrast, when weakly fused persons learn that members of their group are
imperiled, the relative absence of feeling one with the group will ensure that they
experience minimal distress on behalf of the group members. Free of emotional reactions
to the plight of the group, their immediate focus may be on how to insure their own
personal survival. If they have time to reflect, weakly fused participants may go on to ask
themselves if it is morally defensible to murder five ingroup members to save one person
(themselves). They may conclude that it is not. As such, in those relatively rare instances
in which weakly fused individuals do walk down the path to self-sacrifice, they may be
prompted them to do so by carefully reasoned, utilitarian principles. To test these ideas,
we developed two new intragroup versions of the trolley dilemma (Foot, 1967), a widely
used paradigm in the moral judgment literature.
Identity Fusion in Two Intragroup Versions of the Trolley Dilemma
In previous intragroup variations of the trolley dilemma (Swann et al., 2010a),
participants first imagined that they were standing on an overpass above some train tracks.
Down the tracks, they could see that five ingroup members were trapped and were about
to be crushed by a runaway train. As such, participants first learned of the plight of
several ingroup members (activating their social self); only after this did they learn of
Contemplating self-sacrifice 9
their option to save the ingroup members by sacrificing themselves (activating their
personal self). Conceivably, activating the social self prior to the personal self may have
elevated the psychological potency of the social self, causing strongly fused participants
to endorse pro-group behavior to a greater degree than they would have otherwise. In
addition, in this dilemma, the extreme behavior (dying for the ingroup) was always the
pro-group behavior.
To determine if fusion effects were limited by either the order in which the social
self was activated or the relative extremity of the pro-group behavior, we constructed two
new variations of the trolley dilemma that varied these factors. As in previous research
(see Gómez et al., 2011a; Swann et al., 2010a), in one variation participants first learned
that several ingroup members were imperiled (activating their social self) and then
learned that they could save those ingroup members by placing themselves in jeopardy
(activating their personal self and availing them of the possibility of enacting an extreme
pro-group behavior). In a second variation, participants first learned that they themselves
were in jeopardy (activating their personal self) and then learned that they could save
themselves by (surreptitiously) placing several ingroup members in peril (activating their
social self and availing them to the possibility of enacting an extreme anti-group
behavior). In the latter variation, we expected that strongly fused persons would still be
particularly inclined to endorse pro-group action, with weakly fused persons endorsing
the extreme anti-group option of ensuring their own survival by placing several group
members in harm’s way. In short, our new intragroup variations of the trolley dilemma
tested if fusion effects would generalize to scenarios in which the social self was
Contemplating self-sacrifice 10
activated after the personal self and in which there was an option for extreme anti-group
behavior to guarantee personal survival.
We named the first of the new intragroup variations of the trolley dilemma the
summoning the death train scenario. Individual participants learned that a runaway train
was about to crush and kill five ingroup members (citizens of his/her country) unless they
flipped a switch that diverted the train to their own railway track, killing them but leaving
the five ingroup members unharmed. Participants then chose between letting the train
crush the five ingroup members or flipping the switch to sacrifice their own lives.
The second new trolley dilemma was the luring others to their death scenario. It
began by having participants imagine that they were themselves trapped on a railway
track, with a deadly train hurtling their way. Then they learned that a group of five fellow
citizens was 200 meters away, and if participants called them, they would come to help.
Their fellow citizens were unaware, however, that upon arriving, they would become
trapped between the train and the participant, saving the participant but killing them.
Participants then chose between doing nothing and dying versus saving their own lives by
inviting the five ingroup members into a deathtrap.
With these two new variations of the trolley dilemma in hand, we conducted a
series of seven investigations. The first three experiments examined the basic properties
of the “summoning” and “luring” dilemmas. Experiment 1 was designed to provide
information regarding participants’ perceptions of the morality of sacrificing oneself,
versus five ingroup members, in the context of each dilemma. We expected that in
adherence to widely shared social values, most participants would recognize self-sacrifice
as more moral than sacrificing others to save oneself. Experiment 2 asked if, in both
Contemplating self-sacrifice 11
dilemmas, identity fusion would influence endorsement of self-sacrifice, with strongly
fused persons tending to endorse sacrificing themselves to save other ingroup members
while weakly fused participants would not. Experiment 3 tested our assumption that
strongly fused persons endorsed self-sacrifice out of devotion to fellow group members
against the rival hypothesis that they were unconcerned with self-preservation.
In the remaining four experiments, we used the new dilemmas to examine
participants’ conscious thoughts and emotions while they decided whether to
sacrifice themselves or members of their group. We were especially interested in the
processes that underlay the tendency for strongly fused persons to endorse sacrificing
themselves for ingroup members. Insofar as strongly fused persons experience a visceral
feeling of oneness with other group members, when they learned that other group
members were in grave danger, they should experience emotional reactions that would
motivate them to save ingroup members from their plight. Even if they fail to endorse
sacrificing themselves, they should resolve to find another way to save the ingroup
members. In contrast, learning that members of their group are imperiled should produce
relatively little group-related emotional distress among weakly fused persons. Instead,
their initial impulse should be to focus on preserving their own lives. If given enough
time to reflect, some weakly fused participants may become concerned with the
objectively problematic decision of endorsing the murder of five ingroup members to
save themselves. They may accordingly endorse self-sacrifice. We tested these
hypotheses in Experiment 4.
One implication of the foregoing analysis is that curtailing the amount of time that
people have before responding to the dilemma should magnify the differences between
Contemplating self-sacrifice 12
strongly versus weakly fused individuals. That is, if people are informed that members of
their group are imperiled and required to indicate immediately whether they would
sacrifice themselves to save these group members, their emotionally charged, intuited
response should prevail (e.g., Rand, Greene, & Nowak, 2012; Suter & Hertwig, 2011).
Specifically, hurrying participants’ responses should exaggerate the tendency for strongly
fused people to endorse sacrificing themselves for ingroup members and for weakly fused
people to save themselves at the expense of ingroup members. We tested these ideas in
Experiment 5.
A second implication of this analysis is that priming the personal self of
participants should have opposite effects on strongly versus weakly fused persons. That
is, because fused persons have a visceral sense of oneness with the group, priming the
personal self should simultaneously activate both selves. If they then learn that ingroup
members are imperiled, the priming manipulation should amplify their immediate
emotional impulse to sacrifice themselves for the ingroup members. In contrast, because
weakly fused persons lack a visceral sense of oneness with the group, priming the
personal self should activate their personal self only. When they then learn that ingroup
members are imperiled, the priming manipulation should amplify their impulse to focus
on their own personal survival. Their willingness to engage in self-sacrifice should
therefore decline (see also Conway & Gawronski, 2013). We tested these hypotheses in
Experiment 6.
A third implication of this analysis is that the strong relational ties that fused
persons develop toward other ingroup members should prevent them from thinking of
such individuals in utilitarian terms. Note that many people think that the life of even a
Contemplating self-sacrifice 13
single family member is sufficient to justify the sacrifice of their own life. Insofar as
strongly fused persons develop family like ties to fellow group members, they should be
just as inclined to sacrifice their lives for a single member of the ingroup as for multiple
members of the ingroup. In contrast, weakly fused persons may take a relatively
pragmatic view of the value of the lives of ingroup members. Rather than seeing the lives
of each individual ingroup member as sacrosanct, they see the lives of ingroup members
as adding incrementally to the value of the group as a whole. For this reason, when it
comes to sacrificing their lives for ingroup members, weakly fused persons may be
sensitive to utilitarian principles: the more lives at stake, the more inclined they should be
to sacrifice their own life. We tested this hypothesis in Experiment 7.
Experiment 1:
Moral Perceptions of Self-Sacrifice
The goal of this Experiment was to assess perceptions of the morality of each of
the two response options to our intragroup versions of the trolley dilemma. Participants
were randomly assigned to read either the summoning or luring dilemma. They then
indicated the degree to which they perceived sacrificing oneself versus the ingroup
members to be the morally correct course of action.
Method
Participants. Participants were 41 Spanish undergraduate volunteers (27 women,
M age = 34.56, SD = 8.95) enrolled in UNED. All participated online, typically from their
homes.
Procedure. Participants in the summoning condition learned that a runaway train
was headed for five Spaniards (i.e., ingroup members for our Spanish participants) who
Contemplating self-sacrifice 14
would be crushed unless the participant flipped a switch to divert the train to the
participant’s railway track. If diverted, the train would crush the participant but spare the
ingroup members. Participants in the luring condition learned that a runaway train was
headed toward the participant, who would be crushed if no one intervened. They also
learned the participant could call five Spaniards who were 200 meters away. The
Spaniards would come to help, not realizing that upon arrival, they would become
trapped in the path of the train and killed, saving the participant.
After reading the dilemma, participants responded to a binary assessment of
morality. It consisted of a single item which asked: “Which of the two choices (selfsacrifice or survival) is the more moral and ethical choice?”
Participants then completed a continuous measure of perceived morality. It
consisted of three items. Scales were anchored only at the ends and ranged from 1 (to let
the 5 ingroup members die) to 6 (self-sacrifice to save 5 ingroup members). Participants
responded to these items: “The response that is good in general is…”, “The right
response is…”, and “The ethical response is…”. Responses to the three items covaried (α
= .74) so we accordingly summed them.
Results
Binary measure of perceived morality. A chi-square analyses on the choice
participants considered as the more moral was significant, χ² (1) = 26.56, p < .001,
indicating that fully 90.2% of participants choose the self-sacrifice option as the more
moral and ethical. In addition, the type of dilemma had little impact on perceptions of
morality. That is, a binary logistic regression with type of dilemma as the predictor
Contemplating self-sacrifice 15
(summoning vs. luring, effect coding: -1, 1) indicated no effect of dilemma on
perceptions of the extent to which the self-sacrifice was moral and ethical, p = .55.
Continuous measure of perceived morality. We first compared the overall
means for each dilemma against the theoretical midpoint of the scale (3.5). Self-sacrifice
was rated as more moral than allowing the ingroup members to die in both the luring
condition, t (15) = 5.76, p < .001, and in the summoning condition, t (24) = 2.32, p = .029.
An additional analysis, however, revealed an effect of the dilemma on perceptions of
morality. Participants in the luring condition perceived self-sacrifice as more moral than
participants in the summoning condition, t (39) = 2.49, p = .017, M = 5.12, SD = 1.13 vs.
M = 4.12, SD = 1.34, respectively.
Discussion
As expected, the results of Experiment 1 showed that the vast majority of
participants recognized sacrificing oneself to save the ingroup members as morally
superior to sacrificing others to save oneself. Participants were especially inclined to
acknowledge the moral superiority of self-sacrifice in reference to the “luring” dilemma,
probably due to the unsavory nature of the alternative option of tricking others into
sacrificing their lives to save oneself. Indeed, recent research shows that when people
consider committing harm against others, they normatively rely upon a proscriptive
morality that is strict and obligation-based (i.e., to not do harm; Janoff-Bulman, Sheikh,
& Hepp, 2009). One implication of these different perceptions of the dilemmas is that
participants may be more inclined to endorse self-sacrifice in the “luring” than in the
“summoning” dilemma. We tested this possibility in Experiment 2.
Contemplating self-sacrifice 16
The most important goal of Experiment 2, however, was to examine the
relationship of identity fusion to willingness to sacrifice oneself for ingroup members.
We expected that strongly fused participants would be more inclined to endorse selfsacrifice than weakly fused participants. In addition, we expected that these fusion
differences would emerge independent of the perceived morality of endorsing selfsacrifice. For example, we expected that weakly fused individuals would tend to
acknowledge the moral superiority of endorsing self-sacrifice but nevertheless refrain
from endorsing it.
Experiment 2:
Self-Sacrifice for Ingroup Members
Experiment 1 accomplished the first step in validating our new intragroup trolley
dilemmas by showing that most participants acknowledged the moral superiority of selfsacrifice within both dilemmas. The second step involved testing our hypothesis that
identity fusion would predict participant’s responses to these dilemmas.
Method
Participants. Participants were 622 Spanish undergraduate volunteers (357
women, M age = 34.48, SD = 11.22) enrolled in UNED. All participated online.
Procedure. Identity fusion was measured using the 7-item verbal fusion scale
(Gómez et al., 2011a) with reference to the group “My Country,”α = .83. Example items
are “I am one with my country” and “I am strong because of my country”. Respondents
indicated the degree to which each statement reflected their relationship with their
country on scales ranging from 0 (strongly disagree) to 6 (strongly agree). Higher scores
reflected stronger fusion with country.1
Contemplating self-sacrifice 17
Participants then were randomly assigned to the summoning or luring condition.
After indicating their binary response to the dilemma, participants responded to three
items that assessed the perceived morality of their response “I believe that my response is
good in general”, “I think I have chosen the right response”, and “I think that my
response is ethical”. Responses to each question were made on 7-point scales ranging
from 0 (strongly disagree) to 6 (strongly agree). Responses to the three items covaried
(α= .80) so we accordingly summed them.
Results
Response to the dilemma. We used a binary logistic regression to examine the
impact of fusion, dilemma (summoning vs. luring dilemma, effect coding: -1, 1) and the
interaction on self-sacrifice for the ingroup members (in all experiments in this report,
fusion was centered and dichotomous predictors were effect-coded). A main effect of
fusion emerged, B = .67, odds ratio (OR) = 1.96, Wald χ2 = 50.82, p < .001, with strongly
fused participants preferring to sacrifice themselves more than weakly fused participants.
A main effect of the type of dilemma also emerged, B = -.43, odds ratio (OR) =. 65, Wald
χ2 = 23.49, p < .001, such that participants choose to self-sacrifice more in the luring
dilemma than in the summoning dilemma, 49.7% vs. 31.4%. The fusion x dilemma
condition interaction was not significant, B = -.03, odds ratio (OR) =. 97, Wald χ2 = .09, p
= .77.
Perceived morality of one´s response. Although weakly fused persons were
reluctant to endorse self-sacrifice for the ingroup members, we expected that they would
resemble strongly fused persons in recognizing self-sacrifice as more moral than saving
oneself at the cost of others. To test this prediction, we conducted a multiple regression in
Contemplating self-sacrifice 18
which the predictors were fusion, dilemma type, response to the dilemma, and all two
ways and the triple interaction terms as predictors. The criterion was perceived morality.
As expected, only the main effect of the response to the dilemma was significant, b = .50,
t (614) = 9.23, p < .001, such that participants who chose self-sacrifice considered their
response as more moral than those who chose to let the five ingroup members die. All
other terms were not significant, ps > .25.
Discussion
Our findings replicated previous evidence of a link between identity fusion and
willingness to self-sacrifice for the ingroup (Gómez et al., 2011a; Swann et al., 2010a)
with two new intragroup trolley dilemmas. No matter whether the dilemma required that
the participants take action to either sacrifice themselves (i.e. the “summoning” dilemma)
or save themselves (i.e. the “luring” dilemma), strongly fused participants displayed a
stronger preference for self-sacrifice than weakly fused participants.
The perceived morality of the response did not figure prominently in how
participants responded to the dilemma. Not surprisingly, participants who endorsed selfsacrifice indicated that it was the most moral of the two response options. Indeed,
participants who endorsed saving themselves admitted that they had endorsed the less
moral of the two options. Apparently, identity fusion influences whether individuals are
willing to act on what they feel they morally should do.
Our results also showed that the relationships between fusion and self-sacrifice
were similar for each dilemma, as indicated by the lack of an interaction between fusion
and dilemma type. This also means that the effects of identity fusion were similar,
regardless of the order of identity activation (i.e., social before personal, or vice versa) or
Contemplating self-sacrifice 19
the nature of the extreme behavior (dying to save the ingroup members, murdering others
to save oneself).2
The results of the first two experiments suggest that strongly and weakly fused
persons perceived the dilemmas in much the same way. Nevertheless, the relative
willingness of strongly fused persons to endorse self-sacrifice may not have reflected
devotion to the group as we had expected. Instead, strongly fused persons may have
endorsed self-sacrifice because they were less concerned with self-preservation than
weakly fused persons. We addressed these competing explanations of the effects of
fusion in Experiment 3.
Experiment 3:
Is self-sacrifice a sacrifice for strongly fused people?
Method
Participants. Eighty-five Spanish undergraduate students from UNED (69
female, Mage = 31.40, SD = 10.24) participated online for course credit.
Procedure. After learning that the experiment involved responses to moral
dilemmas, participants completed the verbal fusion scale (Gómez et al., 2011a) with
reference to the group “My Country,”α = .85. Then, in counterbalanced order,
participants responded to a self-preservation scale and a devotion to saving ingroup
members scale on 7-point scales ranging from 0 (disagree strongly) to 6 (agree strongly).
The self-preservation scale (α = .70) included 3 items (“The most important thing for me
when I am in danger is to survive,” “I think that the desire to survive is among the most
basic of all motives,” and “I want to live more than anything else”). The devotion to
saving ingroup members scale (α = .70) included three items: “I am devoted to saving
Contemplating self-sacrifice 20
ingroup members,” “I would kill to save members of my group,” and “Protecting the
lives of ingroup members is really important to me.” After completing these scales,
participants were randomly assigned to respond to the luring or the summoning dilemmas.
Results
Relation of fusion to desire for self-preservation and devotion to saving
ingroup members. As expected, fusion was unrelated to desire for self-preservation,
r(83) = .02, p = .85, but related to devotion to saving ingroup members, r(83) = .35, p
= .001.
Response to the dilemmas. Overall, 36.5% of participants chose self-sacrifice.
We used a binary logistic regression to examine the impact of fusion, dilemma
(summoning vs. luring dilemma), and the fusion X condition interaction on self-sacrifice
for ingroup members. The analysis yielded a main effect of fusion, B = 1.89, Wald χ2 =
4.59, OR = 6.61, p = .032, indicating that strongly fused participants were more willing to
sacrifice themselves than weakly fused participants. There was also a main effect of the
dilemma condition, B = -1.28, Wald χ2 = 6.47, OR = .28, p = .01. More participants chose
to self-sacrifice in the luring dilemma than in the summoning dilemma, 50% vs. 23.3%.
The fusion X dilemma condition interaction was not significant, B = -.88, Wald χ2 = 2.62,
OR = .42, p = .10. When we decomposed this (non-significant) interaction, strongly fused
were slightly more willing to self-sacrifice in the luring dilemma, B = 1.01, Wald χ2 =
6.35, OR = 2.75, p = .012, than in the summoning dilemma, B = 2.77, Wald χ2 = 3.87, OR
= 15.91, p = .049, but both values were significant. This finding reinforces our
assumption that fusion has similar effects on responses to the two dilemmas.
Contemplating self-sacrifice 21
Further analyses revealed that self-sacrifice was predicted by devotion to protect
ingroup members, B = .96, Wald χ2 = 11.56, OR = 2.62, p < .001 but not by desire for
self-preservation, p = .98. When fusion and devotion to protect ingroup members were
entered simultaneously as predictors, devotion predicted self-sacrifice, B = .89, Wald χ2 =
9.04, OR = 2.43, p = .003, whereas fusion did not, p = .35, presumably because the
devotion to protect ingroup members scale directly asked about self-sacrifice (correlation
between the devotion to protect ingroup members scale and self-sacrifice is .40, p < .001).
Mediational analyses. We conducted a mediational analysis to determine if the
effect of fusion on response to the dilemma was mediated by devotion. A bootstrapping
test (n boots = 5,000; Preacher & Hayes, 2008) showed that fusion predicted the response
to the dilemma, b = .50, p = .036, and the mediator, devotion, b = .35, p < .001. However,
when the mediator was included in the analyses, the effect of fusion on the choice to the
dilemmas was not significant, p = .35. This suggests that devotion fully mediated the
effects of fusion on endorsement of self-sacrifice, CI = .0859 to .6657.
Discussion
The results of Experiment 3 point to two major conclusions. Most important, the
heightened willingness of strongly fused persons to endorse self-sacrifice cannot be
attributed to a diminished desire for self-preservation, as strongly fused persons were
every bit as committed to self-preservation as weakly fused persons. Furthermore,
devotion to the group – rather than desire for self-preservation – fully mediated the effect
of fusion on endorsement of self-sacrifice.
The second conclusion was that the influence of fusion on responses to the
summoning and luring dilemmas was similar on all outcome measures. This finding,
Contemplating self-sacrifice 22
together with converging evidence from Experiments 1 and 2, provided empirical
justification for using the two dilemmas interchangeably in the remaining experiments in
this article. Experiment 4 utilized the summoning dilemma, and subsequent experiments
alternated between the luring and summoning dilemmas.
Experiment 4:
“Think-Aloud” Protocols as a Window Into Thoughts and Emotions of Participants
If the results of Experiment 3 indicate that devotion to the group guided the
responses of strongly fused participants, they fail to reveal precisely how such devotion
influenced how participants construed the dilemmas. In Experiment 4, we directly
assessed participants’ cognitive and emotional reactions to the summoning dilemma by
presenting them with the dilemma and asking them to think out loud as they decided what
to do. In adopting this “think aloud” paradigm, we acknowledge that self-reports have a
somewhat checkered history within psychology, in general, and social psychology, in
particular (Nisbett & Wilson, 1977). Nevertheless, the most telling critiques of this
approach have focused on retrospective self-reports. These critiques are justified because
when people look back upon their behaviors, a great deal of information is either
completely lost or systematically degraded by numerous sources of bias. Nevertheless,
this problem is vastly diminished when participants report on their thoughts concurrently,
that is, while they make decisions (Ericsson & Simon, 1993). We accordingly examined
participants’ concurrent verbal protocols as they decided how to respond to the dilemma.
Method
Contemplating self-sacrifice 23
Participants. Participants were 293 Spanish undergraduates (158 female, M age =
36.71, SD = 14.10) enrolled at UNED. All completed the experiment via a paper and
pencil questionnaire for course credit.
Procedure. After learning that the experiment involved responses to moral
dilemmas, participants completed some questionnaires including the verbal scale of
fusion with country (α = .89). Participants then read the “summoning the death train”
dilemma. Immediately after reading the dilemma, participants in the think aloud
condition learned that they should verbalize their thoughts about the dilemma as they
were reaching a decision. Participants in the control condition received no such
instructions and instead made their decision quietly. When they had come to a decision,
participants in both groups responded to the dilemma and completed the items tapping
the perceived morality of their response, α = 82.3
Rating the audiotapes. After Swann, Stein-Seroussi, and Giesler (1992), we
sought to develop a coding scheme that would capture all of the thoughts that participants
generated. To this end, we had four undergraduate assistants (who were unfamiliar with
the nature and purposes of our research) listen to participants’ “think aloud” audiotapes.
Assistants identified five distinct rationales that participants offered as they considered
whether they would self-sacrifice for the group or allow the five ingroup members to die.
1. Survival instinct: The speaker asserted that his/her life was the most important
thing and that his/her response was instinctive, natural, and typical because the motive to
survive is the most fundamental human motivation. Example: “The most important
motivation is the instinct to survive.”
Contemplating self-sacrifice 24
2. Distancing from victims. The speaker individuated, derogated, or distanced
victims from the group or from oneself. Example: “I do not know them personally. They
are not my family.”
3. Emotional engagement with the group: The speaker expressed tension, distress
and anxiety regarding the plight of the group members and a near-instantaneous belief
that the morally correct course of action was to sacrifice oneself. Example: “It would be
horrible if they should die and I know without thinking that sacrificing myself is the right
thing to do”.
4. Utilitarianism: The speaker argued that it is better that five as compared to one
Spaniards to survive. Example: “It is better to save five lives than only one, even when
the one life is mine.”
5. Denial: The speaker rejected the binary choice that the experiment presented
and wished for a better alternative. Example: “It should be possible to find a way to save
all five people and me.”
After these five distinct rationales were generated, we began by presenting a new
group of four judges (undergraduates also unfamiliar with the purposes of the research)
with each of the above descriptions of the five categories. A sixth “other” category was
added to accommodate potential responses that did not fit within any of the five
rationales: “If you felt the speaker’s thoughts were not covered by the five categories,
please write down the reason in the space provided” (no judges used this option).
Training consisted of providing judges with an elaborated version of the five categories
to ensure that they understood them. Judges then rated the extent to which participants
verbalized each rationale on a scale ranging from 0 (definitely not the participant´s
Contemplating self-sacrifice 25
rationale) to 4 (definitely the participant´s rationale). To maximize comprehension, we
provided judges with transcripts of the protocols to read as they listened to the audiotapes.
Interjudge reliability (assessed with Cohen’s Kappas) was quite respectable:
survival instinct (.90), distancing victims (.83), emotional engagement with the group
(.92), utilitarianism (.90), and denial (.82). We accordingly averaged all the judges´
response to each item.
Results
Responses to the dilemma. Overall, 22.2% of participants chose to sacrifice
themselves. To determine if our predictor variables influenced responses, we centered
fusion, effect coded the experimental condition (-1, 1), and computed the fusion X
condition interaction (Aiken & West, 1991). With these terms as predictors, we
conducted a binary logistic regression on responses to the dilemma. The analysis yielded
a main effect of fusion, B = .70, Wald χ2 = 19.18, OR = 2.01, p < .001, indicating that
strongly fused participants were more willing to self-sacrifice than weakly fused. No
other effects were significant (p’s > .11).
Rationales for responses to the dilemma. To determine whether participant’s
“think alouds” differed as a function of fusion and response to the dilemma, we
conducted a series of regressions including fusion, response to the dilemma (effect
coding: -1, 1), and the fusion X response interaction on each of the five cognitiveemotional processes. The results of the regressions are shown in Table 1. In all 5
regressions, significant fusion X response interactions qualified main effects of fusion
and response to the dilemma.
Contemplating self-sacrifice 26
To decompose the interactions, we conducted simple slope analyses. As shown in
Figure 1A, among participants who chose to sacrifice themselves, strongly fused persons
were especially likely to verbalize emotional engagement with the group rationales, B
= .60, t (121) = 3.87, p < .001, while weakly fused participants were especially likely to
verbalize utilitarian rationales, B = -.61, t (121) = -3.97, p < .001.4 As shown in Figure
1B, no differences were found for either of these categories among participants who
chose to let five ingroup members die, p’s > .76. Figure 1B also shows that among
participants who chose to let five ingroup members die, strongly fused participants were
especially likely to verbalize denial rationales, B = .54, t (121) = 5.67, p < .001, while
weakly fused participants were especially likely to verbalize survival instinct and
distancing from victim’s rationales, B = -.63, t (121) = -7.70, p < .001, and B = -.65, t
(121) = -7.51, p < .001 respectively. No differences were found on these rationales for
participants who chose self-sacrifice, p’s > .38 (see Figure 1A).
Mediational analyses. We conducted a mediational analysis to determine if the
effect of fusion on response to the dilemma was mediated by emotional engagement
associated with Spaniards. The results are displayed in Figure 2. A bootstrapping test (n
boots = 5,000; Preacher & Hayes, 2008) confirmed our expectation that emotional
engagement with the group partially mediated the effect of fusion on the response to the
dilemma.
Discussion
As in Experiment 2 and 3, strongly fused persons endorsed self-sacrifice for
fellow group members more than weakly fused participants. Mindful of the fact that selfsacrifice was considerably more common among strongly fused participants, we
Contemplating self-sacrifice 27
examined the rationales participants offered as they endorsed self-sacrifice. Strongly
fused endorsers of self-sacrifice responded with rationales that emphasized group-related
emotions, explaining that it was upsetting to them that their group members were in such
a predicament. Many also noted an immediate impulse to forfeit their life to save the lives
of the ingroup members. In contrast, among the few weakly fused participants who
endorsed self-sacrifice, the decision emerged more slowly and was marked by utilitarian
rationales—“Self-sacrifice is called for because the lives of five people are more valuable
than the life of one.” Tellingly, their reasoning was devoid of any reference to attachment
to the group members. Instead, their reasoning focused on minimizing the number of
lives lost.
Strongly and weakly fused participants also differed in the rationales they offered
for failing to endorse self-sacrifice. The mental gymnastics of strongly fused participants
were especially intriguing. Even as they decided to consign the ingroup members to their
deaths by refusing to endorse self-sacrifice, strongly fused participants denied that they
were forsaking their ingroup members. Instead, they insisted that “there must be a
different option!” or “the group members must be saved!” In contrast, when weakly fused
persons endorsed letting ingroup members perish, they insisted that they were, quite
understandably, following their gut survival instincts. They attempted to excuse their
decision to prioritize their own survival over that of group members by attempting to
distance themselves psychologically from the victims.
One way of contrasting the responses of strongly and weakly fused participants is
in terms of the content of their deontological and utilitarian impulses. For strongly fused
persons, their deontological impulse was shaped by emotional distress associated with the
Contemplating self-sacrifice 28
possibility that several group members might die. This impulse often led them to endorse
self-sacrifice. When it did not, their most common follow-up response was to deny the
reality of the dilemma they confronted.
For weakly fused persons, their deontological impulse was shaped by their
survival instinct, an impulse that encouraged them to endorse self-preservation. When
they resisted this impulse, their follow-up response was to engage in utilitarian reasoning,
a process that led a limited number of weakly fused persons to endorse self-sacrifice.
If our participants were indeed motivated by quick, affect-laden, deontological
impulses to either ensure the survival of their group members (strongly fused) or insure
their own survival (weakly fused), then manipulations that are designed to amplify
deontological impulses should exaggerate the characteristic responses of participants. To
test this possibility, in Experiment 5 we hurried participants’ responses, a technique that
research on moral decision making has shown to amplify deontological responding (Suter
& Hertwig, 2011).
Experiment 5: Amplifying Fusion Effects Through Time Pressure
In this experiment, we hurried the responses of some participants using a timepressure manipulation. We predicted that pressure to respond quickly would exaggerate
the tendency for strongly fused persons to self-sacrifice for the group relative to
participants who were not under time pressure. In contrast, we predicted that pressure to
respond quickly would exaggerate the tendency for strongly fused persons to refrain from
self-sacrifice for the group relative to participants who were not under time pressure. We
tested these predictions using the luring dilemma.
Method
Contemplating self-sacrifice 29
Participants. Participants were 607 students at UNED who participated online for
course credit. Nineteen participants were not included in the analysis because they failed
to correctly identify the nationality of the five members in the track or failed to respond
to this item. After Rand et al. (2012), we excluded from the analyses participants in the
time-pressure condition who took longer than 10 seconds (N = 52) and participants in the
control condition who took less than 10 seconds (N = 100). This left a sample of 436
participants (242 women, M age = 33.90, SD = 11.08) for analysis. Although restoring
these participants to the analyses weakened our findings slightly, all significant effects
remained significant.5
Procedure. After learning that the experiment involved responses to moral
dilemmas, participants responded to some questionnaires that included the verbal scale of
fusion with country, α = .81. Participants were randomly assigned to one of two
experimental conditions. In the time-pressure condition, participants received instructions
to respond to the dilemma in no more than 10 seconds. In the control condition,
participants received no instructions about timing. Participants then read the luring
dilemma. Afterwards, they completed the same outcome measures used in the previous
experiments.
Results
Response to the dilemma. Overall, 59.6% of participants chose to self-sacrifice.
We used a binary logistic regression to examine the impact of fusion (centered), dilemma
condition (control versus time-pressure; effect coding: -1, 1), and the fusion X condition
interaction on response to the dilemma. A main effect of fusion emerged, B = 1.06, OR =
2.89, Wald χ2 = 51.06, p < .001, indicating that strongly fused participants preferred to
Contemplating self-sacrifice 30
sacrifice themselves for the ingroup members. This main effect was qualified by a fusion
x time pressure interaction, B = .78, OR = 2.18, Wald χ2 = 27.65, p < .001. As seen in
Figure 3, strongly fused participants were more willing to self-sacrifice in the timepressure condition than in the control condition, B = .51, OR = 1.67, Wald χ2 = 11.76, p
< .001. In contrast, weakly fused participants were less willing to self-sacrifice in the
time-pressure condition than in the control condition, B = -.73, OR = .48, Wald χ2 =26.14,
p < .001. No other effects were significant, p’s > .25.
Discussion
Our findings confirmed our expectation that strongly versus weakly fused people
would have very different responses to the dilemma when given little time to consider the
situation. In particular, encouraging participants to respond to the dilemma quickly
increased the tendency for participants to “go with their gut.” For strongly fused persons,
their gut told them to save their group members; for weakly fused persons, their gut told
them to save themselves. Presumably, this pattern reflected a tendency for time pressure
to amplify participants’ unique deontological responses to the dilemma (Suter & Hertwig,
2011).
Time pressure is surely not the only way to engage the visceral sense of oneness
that strongly fused persons have with their group, however. Alternatively, focusing
attention on either the personal or social self should amplify the tendency of fused
persons to translate their pro-group sentiments into action. In contrast, because weakly
fused persons lack a sense of oneness between the personal and social self, the only way
to amplify their tendency to enact pro-group action is to focus attention on the social self;
attention to the personal self should have no such effects. In the next experiment, we test
Contemplating self-sacrifice 31
this reasoning by examining participants’ reactions to manipulations that prime either
their personal or social self.
Experiment 6: Activating the Personal vs. Social Self
To test our hypotheses, in Experiment 6 we examined the effects of priming either
the personal or social self relative to a base-line control condition in which participants
received no prime whatsoever. Insofar as strongly fused persons experience a visceral
sense of oneness with the group, priming the personal self should simultaneously activate
their social and personal self, resulting in elevated levels of self-sacrifice for the group in
both conditions relative to controls. Insofar as weakly fused persons lack a visceral sense
of oneness with the group, priming the social self should produce elevated levels of selfsacrifice relative to those in personal-self-activation condition and controls.
Method
Participants. Participants were 572 Spanish undergraduates (377 women, M age =
33.21, SD = 9.83) enrolled at UNED. All participated online in exchange for course credit.
Procedure. After learning that the Experiment involved responses to moral
dilemmas, participants completed the fusion with one’s country scale, α = .85.
Participants were then randomly assigned to one of the three conditions. Participants in
the personal self activation condition were asked to think about their personal selves (i.e.,
what they are personally like, things they personally like, words that best describe
personal features). Participants in the social self activation condition were asked to think
about themselves as Spaniards (i.e., what they and other members of their country are
like, the things they like as Spaniards, and words that best describe their Spanish features).
Participants in the control condition were asked to objectively describe the place where
Contemplating self-sacrifice 32
they were answering the questionnaire (e.g., their apartment). After the manipulation,
participants read and responded to the summoning dilemma.
Results
Response to the dilemma. Overall, 29.7% endorsed the self-sacrifice option. We
followed the same analytic strategy as Swann et al. (2009, Experiment 3) to determine if
fusion and identity activation condition interactively predicted dilemma responses.
Fusion was first centered. Then, two orthogonal contrasts were created for the identity
activation variable. The first orthogonal contrast compared the control condition with the
mean of personal and social self activation condition (-2, 1, 1). The second orthogonal
contrast compared the personal self activation condition with the social self activation
condition (0, 1, -1). Then, in a binary logistic regression, fusion, identity activation
condition (orthogonally coded), and both two-way interactions were entered as predictors
of the response to the dilemma.
As predicted, the regression showed an interaction between the first orthogonal
contrast and fusion, B = .29, Wald χ2 = 10.17, OR = 1.33, p = .001, indicating that the
difference between the control condition and the mean of personal and social self
activation conditions varied as a function of fusion. As can be seen in Figure 4, among
strongly fused participants, those in both the personal and social self activation conditions
were more willing to self-sacrifice than control participants, B = .40, Wald χ2 = 18.79,
OR = 1.50, p < .001, but no such pattern emerged among weakly fused persons, B = .12,
Wald χ2 = .92, OR = 1.13, p = .34.
The analysis also revealed an interaction between the second orthogonal contrast
and fusion, B = -.97, Wald χ2 = 27.41, OR = .38, p < .001, indicating that the difference
Contemplating self-sacrifice 33
between the personal and the social self activation conditions varied as a function of
fusion. As shown in Figure 4, strongly fused participants responded similarly in the
personal self and the social self activation conditions B = -.13, Wald χ2 = .67, OR = .88, p
= .41. Weakly fused participants expressed more willingness to self-sacrifice for the
ingroup members in the social self activation condition than in the personal self
activation condition B = 1.55, Wald χ2 = 40.61, OR = 4.73, p < .001.
The foregoing interactions qualified two main effects. First, there was a main
effect of fusion, B = 1.07, Wald χ2 = 57.65, OR = 2.91, p < .001 such that strongly fused
persons expressed more willingness to self-sacrifice than weakly fused persons. Second,
there was a main effect of the second orthogonal contrast, B = .77, Wald χ2 = 22.07, OR =
2.16, p < .001, indicating that self-sacrifice was greater in the social self activation
condition than in the personal identity activation condition. No other effect was
significant (p = .13).
Discussion
Consistent with our assumption that strongly fused persons have a visceral sense
of oneness with the group, our findings revealed that strongly fused participants were
particularly inclined to endorse self-sacrifice when either their personal or social selves
were activated. In contrast, weakly fused participants were especially inclined to endorse
self-sacrifice after their social selves were activated but not after their personal selves
were activated. This pattern of results also supports our assumption that strongly fused
persons experience close relational ties with fellow group members. That is, insofar as
fused individuals develop strong, family like ties between themselves and other group
Contemplating self-sacrifice 34
members, it is not surprising that activating their personal selves augmented their desire
to protect the individuals with whom they were so strongly aligned.
The strong relational, family like ties that fused persons presumably develop
toward other ingroup members have a further implication. To the extent that people
experience a sense of oneness with the group, they may be just as inclined to sacrifice
their lives for a single member of the ingroup as for multiple members of the ingroup.
After all, if the death of a “family” member is equivalent to my own death, then the death
of five family members is likewise my own death. In contrast, weakly fused persons may
take a more emotionally detached, analytic view of the value of the lives of ingroup
members. Rather than seeing the lives of individual ingroup members as equivalent to
their own death, they see the lives of each ingroup member as adding incrementally to the
group category. For this reason, when it comes to sacrificing their lives for ingroup
members, weakly fused persons may be sensitive to utilitarian principles: the more lives
that are in danger, the more inclined they should be to sacrifice their own life. We tested
this hypothesis in Experiment 7.
Experiment 7: For Whom is Self-Sacrifice Utilitarian?
Method
Participants. Participants were 1368 Spanish undergraduates at UNED (757
female, M age = 35.14, SD = 11.15). All participated online for course credit.
Procedure. After learning that the experiment involved responses to moral
dilemmas, participants completed the verbal scale of fusion with country, α = 81.
Participants were randomly assigned to read one of two versions of the luring dilemma.
Contemplating self-sacrifice 35
In one version, there were five ingroup members whom one could lure to their deaths to
save oneself. In the other version, there was only one ingroup member to lure.
Results
Response to the dilemma. Overall, 62.2% of participants chose self-sacrifice.
We used a binary logistic regression to examine the impact of fusion (centered), dilemma
condition (five versus one other ingroup member to lure; effect coded -1, 1), and the
fusion X condition interaction on the response to the dilemma. The analysis yielded a
main effect of fusion, B = .53, Wald χ2 = 63.55, OR = 1.70, p < .001, indicating that
strongly fused participants were more willing to self-sacrifice. There was also a main
effect of the experimental condition, B = -.13, Wald χ2 = 5.16, OR = .87, p = .023. More
participants chose to self-sacrifice for five ingroup members than for one ingroup
member, 66.8% vs. 60.6%. These main effects were qualified by a fusion X condition
interaction, B = .25, Wald χ2 = 14.43, OR = 1.76, p < .002. As see in Figure 5, simple
slope analyses showed that weakly fused participants were more willing to sacrifice for
five ingroup members than for one ingroup member, B = -.39, Wald χ2 = 19.83, OR = .68,
p < .001. However, strongly fused were equally willing to sacrifice for one ingroup
member compared to five ingroup members.
Discussion
The results of Experiment 7 revealed that strongly fused participants were equally
willing to self-sacrifice for one as for five ingroup members. Such individuals therefore
appeared to follow deontological principles (e.g., ‘do no harm to any group member’), in
that the life of even a single ingroup member was sufficient to justify sacrificing one’s
own life. In contrast, weakly fused appeared to follow utilitarian principles when
Contemplating self-sacrifice 36
deciding to sacrifice themselves for the group, expressing more willingness to selfsacrifice for five than for one ingroup member.
General Discussion
Although many ridiculed Fan Meizhong for leaving his students in the lurch when
an earthquake struck, our findings suggest that Fan is not alone in his unwillingness to
risk his life to save others. Our participants responded to a pair of intragroup moral
dilemmas in which it was apparent to most people that the morally correct response was
to sacrifice one’s life to save several people. Nevertheless, about half of our participants
endorsed saving their own life over the morally correct response, even when saving
themselves came at the expense of the lives of five innocents. Hence, although nearly all
of our participants knew what was right, only those who were strongly fused endorsed
translating their moral knowledge into doing what was right. This raises an intriguing
question: what were our participants thinking and feeling? More specifically, what were
the cognitive and emotional processes responsible for the moral resolve of fused
individuals?
We began with the assumption that the state of fusion is distinguished by a feeling
of oneness with the group and the formation of close, family-like, relational ties to other
group members. The responses of our participants to the dilemmas confirmed this
conceptualization. For example, although strongly fused persons were just as concerned
with self-preservation as their weakly fused counterparts, strongly fused persons were
unique in their devotion to saving members of their group. Further, when strongly fused
persons learned that members of their group were imperiled, they reported experiencing
emotions similar to what they would experience were they themselves imperiled. These
Contemplating self-sacrifice 37
emotional reactions, in turn, mediated the effects of fusion on the endorsement of
sacrificing oneself for the ingroup.
Among weakly fused participants, the (less frequently traveled) pathway to selfsacrifice was quite different. Lacking strong relational ties toward their compatriots,
when they pondered sacrificing themselves, weakly fused participants initially expressed
concerns with their own survival and spoke of their threatened ingroup members in
distant, unemotional, terms. Later, they shifted to a consideration of utilitarian
considerations, such as the number of potential victims who would die in the dilemma.
As a result, they were more apt to endorse self-sacrifice for five victims as compared to
one.
Strength of identity fusion was also related to the reasons for refraining from selfsacrifice. When strongly fused participants declined to endorse self-sacrifice, they
denied that they were forsaking the ingroup, clinging instead to the illusory idea that
there must be a way to save both the ingroup members and themselves. In contrast,
weakly fused participants distanced themselves from the group members and focused on
themselves. They insisted that their refusal to endorse self-sacrifice was only natural
given their possession of a quintessentially human survival motive.
We expected that one of the clearest differences between fused and non-fused
participants would be their deontological impulses. On the one hand, we anticipated that
upon learning that members of their group were in jeopardy, strongly fused persons
would mentally transport themselves into the situation facing their compatriots and
experience the group members’ emotions as their own. On the other hand, we expected
that the threat that the dilemma posed to their personal survival would dominate the
Contemplating self-sacrifice 38
immediate reactions of weakly fused persons. For these reasons, we predicted that
requiring participants to hurry their response to the dilemma would amplify the divergent
deontological impulses of strongly and weakly fused participants. Our results supported
this prediction. When we placed participants under time pressure by requiring that they
respond to the dilemma quickly, strongly fused participants were especially inclined to
endorse self-sacrifice and weakly fused participants were especially disinclined to
endorse self-sacrifice. This suggests that when people feel compelled to act quickly
because they are under threat they may be especially likely to translate their feelings of
fusion with the group into extreme pro-group behavior.
This difference in the deontological impulses of strongly and weakly fused
persons to the dilemmas implies that activating the source of such impulses – their
personal selves – should have very different consequences. To test this prediction, we
primed the social and personal selves of participants. Not surprisingly, priming their
social selves increased endorsement of self-sacrifice among both strongly and weakly
fused participants. More interestingly, priming the personal self increased endorsement of
self-sacrifice among strongly fused participants but further reduced endorsement of selfsacrifice among weakly fused participants.6 This evidence that activating the personal
selves of strongly fused persons functioned in the same way as activating their social
selves provides further evidence of the visceral sense of oneness that strongly fused
participants experience with the group.
In summary, in the experiments reported here, we attempted to illuminate directly
the nature of the emotional and cognitive processes that cause strongly fused persons to
embrace self-sacrifice for the group and weakly fused persons to embrace their own
Contemplating self-sacrifice 39
survival. Our findings indicated that strongly and weakly fused persons differed not only
in how often they endorsed self-sacrifice to save others, but in the psychological pathway
they followed to this destination. When they recognized that their group members were in
danger, strongly fused persons immediately felt as if they themselves were in danger and
it did not matter whether one or several ingroup members were threatened. At the same
time, weakly fused persons focused on their own survival and endorsed the sacrifice of
group members. On those relatively rare occasions that weakly fused people winded up
endorsing self-sacrifice, they were inspired to do so by a cool calculus that focused on
utilitarian concerns such as the impact of their decision to the number of people who
would live or die.
Links to recent work on moral reasoning
Over the last two decades, theorists and researchers have increasingly
acknowledged dual process models in which controlled and automatic processes play
complementary roles in moral decision making (e.g., Greene et al., 2008; Lombrozo,
2009). Our findings revealed instances of both automatic and controlled processing in the
moral decision-making of our participants. For example, the relatively automatic,
deontological response of strongly fused participants was to save the lives of group
members who were imperiled. In contrast, the relatively automatic, deontological
response of weakly fused participants was to save themselves.
Additional reflection took strongly and weakly fused participants in very different
directions. On the path to deciding to save themselves at the expense of the group,
strongly fused participants struggled (without success) to develop a rationale that would
allow them to simultaneously save both themselves and the group. On the path to
Contemplating self-sacrifice 40
deciding to save the group, weakly fused participants noted that sacrificing themselves
would result in fewer people dying than allowing five ingroup members to perish.
Apparently, strongly versus weakly fused persons diverged in both their spontaneous and
reasoned reactions to the dilemmas.
Our evidence of the impact of the immediate, emotional responses of fused persons
on their moral decision-making has some parallels in the recent literature on sacred
values and culture of honor. In studies of extremists in the Middle East, for example,
researchers have used the label “devoted actors” to refer to individuals whose decisions
are determined by “sacred values” that turn on a rule-bound, emotionally arousing moral
logic (Sheikh, Ginges, Coman & Atran, 2012). The rationales underlying the pro-group
behavior of these devoted actors are irrational in that they are insensitive to quantity and
abhor material tradeoffs (Atran & Ginges, 2012; Ginges, Atran, Medin, & Shikaki, 2007;
Ginges, Atran, Sachdeva, & Medin, 2011; see also Skitka, 2002, 2010, for parallel
findings with regard to ‘moral convictions’). Similarly, individuals who adhere to the
principles of culture of honor react to perceived slights to in-group members with
“illogical” levels of retribution that far exceed what could be justified from an objective
standpoint (e.g., Cohen & Nisbett, 1994; Cohen, Nisbett, Bowdle, & Schwarz, 1996).
Our data suggest that identify fusion may play a role in the emotional responses of such
individuals. Witness, for example, that fused persons became highly emotional when they
learned of the plight of members of their group. Moreover, the deontological impulses of
strongly fused persons rendered them uniquely insensitive to the number of potential
victims of the runaway trolley in Experiment 7.
Contemplating self-sacrifice 41
Such tantalizing links between our research and recent work on moral decisionmaking notwithstanding, one limitation of our work is the use of hypothetical dilemmas
(for a review of research on identity fusion that employs alternative methodologies, see
Swann et al., 2012). It is reassuring that recent work has demonstrated links between
responses to moral dilemmas and theoretically relevant variables such as empathic
concern, religiosity, perspective taking and moral identity (Conway & Gawronski, 2012;
see also Bartels, 2008; Schnall, Haidty, Clore, & Jordon, 2008). Nevertheless, future
work is needed to better understand relationships between in situ moral reasoning about
self-sacrifice and actual engagement in self-sacrificial behavior. An interesting step along
these lines includes interviews of terrorists that have highlighted the close associations
between steadfast group commitment, moral justifications, and follow-through of
extremist attacks (Atran, 2010; Stern, 2003). Much remains to be done, however.
Conclusions
Most people acknowledge that when several members of their group are imperiled,
they should try to save them, even if it means risking their own life. Yet when asked what
they would actually do if confronted with this dilemma, many people said that they
would selfishly opt to save themselves. In fact, even when the only way to save their own
life was to lure five unsuspecting compatriots into a deathtrap, roughly half of our
participants endorsed this morally bankrupt option. Not so for people who felt strongly
fused to their group. The willingness of strongly fused persons to sacrifice themselves to
save members of their group appear to be mediated by the immediate, emotional
reactions that they experienced when they imagined that their group members were in
danger. For strongly fused persons, it was almost as if they were themselves, or a family
Contemplating self-sacrifice 42
member, were in danger. These emotional reactions caused them to do what most people
naively believe they would do: endorse translating a sense of what they morally should
do into a decision about what they would do.
Contemplating self-sacrifice 43
Footnotes
1
In this investigation and all of the experiments in the paper we also included Mael and
Ashforth’s (1992) group identification scale and conducted preliminary analyses
including it as a predictor. As in previous research on identity fusion (see Swann et al.,
2012), identification effects were always weaker than fusion effects and never qualified
the effects of fusion. We accordingly deleted identification from the primary analyses.
2
To test the group specificity of fusion effects, we also conducted an exact replication of
Experiment 2 except we substituted 5 outgroup members (i.e., “foreigners”) for 5 ingroup
members (i.e., Spaniards) in both dilemmas. As expected, fusion with Spain had no
impact on endorsement of self-sacrifice in either dilemma. This result replicates previous
evidence that fusion with Spain had no influence on willingness to sacrifice the self for
members of other countries (Swann et al., 2010a). In addition, fusion was unrelated to
perceptions of morality in both dilemmas. Consistent with Experiment 2, the only
significant predictor of morality perceptions was one’s response to the dilemma. Details
of this experiment are available from the first author.
3
In this experiment as well as experiments 5-7, we included a measure of perceived
morality. In every instance, analyses revealed only a main effect of the response to the
dilemma, B = .33, t (566) = 7.50, p < .001, such that participants who chose self-sacrifice
considered their behavior as more moral than those who chose to let the five ingroup
members die. To minimize redundancy, we refrained from reporting this variable in
experiments 5-7.
Contemplating self-sacrifice 44
4
These results are based on the explanations provided by the 134 participants in the think
aloud condition. Of these 134, 9 did not say anything during the think aloud period,
yielding a total N of 125. We included 3 predictors in the regression: fusion, choice, and
fusion x choice interaction, yielding 121 degrees of freedom.
5
When we restored all participants for the analyses, the sample was N = 588 (336 women,
M age = 34.01, SD = 11.10). Alpha for the verbal scale of fusion, α = .83. On the response
to the dilemma, overall, 61.4% of participants chose to self-sacrifice. The binary logistic
regression on response to the dilemma yielded a fusion x time-pressure interaction, B
= .57, OR = 1.76, Wald χ2 = 24.13, p < .001. Strongly fused participants were more
willing to self-sacrifice in the time-pressure condition than in the control condition, B
= .36, OR = 1.43, Wald χ2 = 7.79, p = .005. In contrast, weakly fused participants were
less willing to self-sacrifice in the time-pressure condition than in the control condition, B
= -.66, OR = .51, Wald χ2 = 29.88, p < .001. This interaction qualified a main effect of
fusion, B = .90, OR = 2.46, Wald χ2 = 61.03, p < .001, wherein strongly fused participants
preferred to sacrifice themselves for the group to a greater degree than weakly fused
participants. No other effects were significant, p’s > .09.
5
This finding might seem to clash with evidence that activating the personal self did not
diminish the tendency for non-fused participants to endorse dying for the group (Swann
et al., 2009). This likely reflects procedural differences. Unlike the earlier experiment, in
this dilemma the personal self was activated by informing participants that they were
going to be crushed by the train before referencing the group members. This may have
reinforced the impact of the personal-self-activation manipulation. In the earlier
Contemplating self-sacrifice 45
experiment, the group members were referenced first, which may have muted the effect
of activating the personal self.
Antecedents of self-sacrifice 46
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting
interactions. Newbury Park. CA: Sage.
Aron, A., Aron, E. N., Tudor, M., & Nelson, G. (1991). Close relationships as including
other in the self. Journal of Personality and Social Psychology, 60, 241-253.
Atran, S. (2010). Talking to the enemy: Faith, brotherhood, and the (un) making of
terrorists. New York, NY: Harper Collins.
Atran, S., & Ginges, J. (2012). Religious and sacred imperatives in human
conflict. Science, 336, 855-857.
Bartels, D. M. (2008). Principled moral sentiment and the flexibility of moral
judgment and decision making. Cognition, 108, 381-417.
Billig, M., & Tajfel, H. (1973). Social categorization and similarity in intergroup
behaviour. European Journal of Social Psychology, 3, 27-52.
Blair, R. J. R. (1995). A cognitive developmental approach to morality: Investigating the
psychopath. Cognition, 57, 1-29.
Buhrmester, M. D., Gómez, A., Brooks, M. L., Morales, J. F., Fernández, S., & Swann,
W. B. Jr. (2012). My group's fate is my fate: Identity-fused Americans and
Spaniards link personal life quality to outcome of 2008 elections. Basic and
Applied Social Psychology, 34, 527-533.
Chaiken, S., & Trope, Y. (Eds.). (1999). Dual process theories in social psychology. New
York: Guilford Press.
Cohen, D., & Nisbett, R. E. (1994). Self-protection and the culture of honor: Explaining
southern violence. Personality and Social Psychology Bulletin, 20, 551-567.
Antecedents of self-sacrifice 47
Cohen, D., Nisbett, R. E., Bowdle, B. F., & Schwarz, N. (1996). Insult, aggression, and
the southern culture of honor: An" experimental ethnography.".Journal of
Personality and Social Psychology, 70, 945-960.
Conway, P., & Gawronski, B. (2013). Deontological and utilitarian inclinations in moral
decision making: A process dissociation approach. Journal of Personality and
Social Psychology, 104, 216-235.
Ellemers, N., Spears, R., & Doosje, B. (1997). Sticking together or falling apart: In-group
identification as a psychological determinant of group commitment versus
individual mobility. Journal of Personality and Social Psychology, 72, 617-626.
Ericsson, K. A., & Simon, H. A. (1993). Protocol analysis: Verbal reports as data
(revised edition). Cambridge, MA: MIT Press.
Foot, P. (1967). The problem of abortion and the doctrine of the double effect. Oxford
Review, 5, 5-15.
Ginges, J., Atran, S., Medin, D., & Shikaki, K. (2007). Sacred bounds on rational
resolution of violent political conflict. Proceedings of the National Academy of
Sciences, 104, 7357-7360.
Ginges, J., Atran, S., Sachdeva, S., & Medin, D. (2011). Psychology out of the
laboratory: The challenge of violent extremism. American Psychologist, 66, 507519.
Gómez, Á., Brooks, M., Buhrmester, M., Vázquez, A., Jetten, J., & Swann, W. B.
(2011a). On the nature of identify fusion: Insights into the construct and a new
measure. Journal of Personality and Social Psychology, 100, 918-933.
Antecedents of self-sacrifice 48
Gómez, Á., Morales, J. F., Hart, S., Vázquez, A., & Swann, W. B., Jr. (2011b). Rejected
and excluded forevermore, but even more devoted: Irrevocable ostracism
intensifies loyalty to the group among identity highly fused persons. Personality
and Social Psychology Bulletin, 37, 1574-1586.
Greene, J. D. (2007). The secret joke of Kant’s soul. In W. Sinnott-Armstrong (Ed.),
Moral Psychology (Vol. 3, pp. 35-80). Cambridge, MA: MIT Press.
Greene, J. D. (2009). The cognitive neuroscience of moral judgment. In M. Gazzaniga
(Ed.), The Cognitive Neurosciences (4th ed., pp. 987-999). Cambridge, MA: MIT
Press.
Greene, J. D., Morelli, S. A., Lowenberg, K., Nystrom, L. E., & Cohen, J. D. (2008).
Cognitive load selectively interferes with utilitarian moral judgment.
Cognition, 107, 1144-1154.
Haidt, J. (2001). The emotional dog and its rational tail: a social intuitionist approach to
moral judgment. Psychological review, 108, 814-834.
Jacoby, L. L. (1991). A process dissociation framework: Separating automatic from
intentional uses of memory. Journal of Memory and Language, 30, 513-541.
Janoff-Bulman, R., Sheikh, S., & Hepp, S. (2009). Proscriptive versus prescriptive
morality: Two faces of moral regulation. Journal of Personality and Social
Psychology, 96, 521-537.
Kohlberg, L. (1969). Stage and sequence: The cognitive-developmental approach to
socialization. In D. A. Goslin (Ed.), Handbook of socialization theory and
research (pp. 347-480). New York: Rand McNally.
Antecedents of self-sacrifice 49
Lombrozo, T. (2009). The role of moral commitments in moral judgment. Cognitive
Science 33, 273-286.
Mael, F., & Ashforth, B. (1992). Alumni and their alma maters: A partial test of the
reformulated model of organizational identification. Journal of Organizational
Behavior, 13, 103-123.
Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition,
emotion, and motivation. Psychological review, 98, 224-253.
Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports
on mental processes. Psychological Review, 84, 231-259.
Pizarro, D. A., & Salovey, P. (2002). On being and becoming a good person: the role of
emotional intelligence in moral development and behavior. In J. Aronson (ed.),
Improving academic achievement: Impact of psychological factors on education
(pp. 247–266). San Diego: Academic Press.
Posner, M. I., & Snyder, C. R. R. (1975). Facilitation and inhibition in the processing of
signals. In P. M. A. Rabbitt and S. Domic (Eds.), Attention and performance V
(pp. 669-682). New York: Academic Press.
Postmes, T., Haslam, S. A., & Jans, L. (2013). A single-item measure of social
identification: Reliability, validity, and utility. British Journal of Social
Psychology, 52, 597-617.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for
assessing and comparing indirect effects in multiple mediator models. Behavior
research methods, 40, 879-891.
Antecedents of self-sacrifice 50
Rand, D. G., Greene, J. D., & Nowak, M. A. (2012). Spontaneous giving and calculated
greed. Nature, 489, 427-430.
Schnall, S., Haidt, J., Clore, G. L., & Jordan, A. H. (2008). Disgust as embodied moral
judgment. Personality and Social Psychology Bulletin, 34, 1096-1109.
Sheikh, H., Ginges, J., Coman, A., & Atran, S. (2012). Religion, group threat and sacred
values. Judgment and Decision Making, 7, 110-118.
Skitka, L. J. (2002). Do the means always justify the ends, or do the ends sometimes
justify the means? A value protection model of justice reasoning. Personality and
Social Psychology Bulletin, 28, 588-597.
Skitka, L. J. (2010). The psychology of moral conviction. Social and Personality
Psychology Compass, 4, 267-281.
Smith, E. R., Seger, C. R., & Mackie, D. M. (2007). Can emotions be truly group level?
Evidence regarding four conceptual criteria. Journal of Personality and Social
Psychology, 93, 431-446.
Spencer, R. (2008, June 2). China earthquake: Teacher admits leaving pupils behind as he
fled Chinese earthquake. The telegraph. Retrieved from
http://www.telegraph.co.uk
Stern, J. (2003). Terror in the name of God: Why religious militants kill. New York: Ecco.
Suter, R. S., & Hertwig, R. (2011). Time and moral judgment. Cognition, 119, 454-458.
Swann, W. B., Jr., Buhrmester, M., Gómez, Á., Jetten, J., Bastian, B., Vázquez, A. et al.
(in press). What makes a group worth dying for? Identity fusion fosters
perception of familial ties, promoting self-sacrifice. Journal of Personality and
Social Psychology.
Antecedents of self-sacrifice 51
Swann, W. B., Jr., Gómez, Á., Dovidio, J. F., Hart, S., & Jetten, J. (2010a). Dying and
killing for one’s group: Identity fusion moderates responses to intergroup versions
of the trolley problem. Psychological Science, 21, 1176-1183.
Swann, W. B., Jr., Gómez, Á., Huici, C., Morales, F., & Hixon, J. G. (2010b). Identity
fusion and self-sacrifice: Arousal as catalyst of progroup fighting, dying, and
helping behavior. Journal of Personality and Social Psychology, 99, 824-841.
Swann, W. B., Jr., Gómez, Á., Seyle, C. D., Morales, J. F., & Huici, C. (2009). Identity
fusion: The interplay of personal and social identities in extreme group behavior.
Journal of Personality and Social Psychology, 96, 995-1011.
Swann, W. B. Jr., Jetten, J., Gómez, Á., Whitehouse, H., & Bastian, B. (2012). When
group membership gets personal: A theory of identity fusion. Psychological
Review, 19, 441-456.
Swann, W. B., Stein-Seroussi, A., & Giesler, R. B. (1992). Why people selfverify. Journal of Personality and Social Psychology, 62, 392-401.
Tajfel, H., & Turner, J. (1979). An integrative theory of intergroup conflict. In W Austin
& S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33-47).
Monterey, CA: Brooks/Cole.
Turiel, E. (1983). The development of social knowledge: Morality and convention.
Cambridge, England: Cambridge University Press.
Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S.
(1987). Rediscovering the social group: A self-categorization theory. Oxford, UK:
Basil Blackwell.
Antecedents of self-sacrifice 52
Turner, J. C., Sachdev, I., & Hogg, M. A. (1983). Social categorization, interpersonal
attraction and group formation. British Journal of Social Psychology, 22, 227-239.
Antecedents of self-sacrifice 53
Table 1. Predictors of Thoughts and Feelings in Experiment 4
Main effect of the decision to
Main effect of identity fusion
Fusion x decision to the
the dilemma
dilemma interaction
B
b
Survival instinct
-.70***
-.33***
-.31***
-.66***
-.37*** -.31***
.45**
.25**
.21**
Distancing victims
-.44**
-.22**
-.21**
-.59***
-.36*** -.30***
.48***
.28***
.24***
Emo. engagement
.48***
.37***
.34***
.34***
.31***
.31**
.27**
.24**
Utilitarianism
.65***
.61***
.57***
-.27***
-.30*** -.25***
Denial
-.28*
.-16*
.34**
-.17*
sr
B
b
.25**
sr
.26***
.21**
B
sr
-.26***
-.29***
-.25***
-.41**
-.28**
-.25**
Note. B = raw regression coefficient; b = standardized regression coefficient; sr =
semipartial correlation. * p < .05, **p < .01, ***p < .001.
b
Antecedents of self-sacrifice 54
Figure Captions
Figure 1A. Thoughts and feelings of participants who decided to sacrifice themselves for
the group in Experiment 4.
Figure 1B. Thoughts and feelings of participants who chose to allow five ingroup
members to die in Experiment 4.
Figure 2. Emotional engagement mediates the effect of identity fusion on self-sacrifice in
Experiment 4.
Figure 3. Self-sacrifice as a function of identity fusion and time-pressure in Experiment 5.
Figure 4. Self-sacrifice as a function of identity fusion and identity activation in
Experiment 6.
Figure 5. Self-sacrifice as a function of identity fusion and number of ingroup members
in Experiment 7.
Antecedents of self-sacrifice 55
Figure 1A.
Definitely the
participant´s
rationale
4
3
2
Weakly fused
Strongly fused
1
0
Survival
Definitely not
the participant´s
rationale
Distancing
Emot/Moral Utilitarianism
engagement
Denial
Antecedents of self-sacrifice 56
Figure 1B.
Definitely the
participant´s
rationale
4
3
2
Weakly fused
Strongly fused
1
0
Survival
Definitely not
the participant´s
rationale
Distancing
Emot/Moral Utilitarianism
engagement
Denial
Antecedents of self-sacrifice 57
Figure 2.
.31***
Emotional
Engagement
.93***
(.79***)
Identity Fusion
.52 (p = .07)
95% Confidence Interval [.0607, .6162]
95% Confidence Interval [0.0607, 0.6162]
Self-sacrifice vs. Not
self-sacrifice
Antecedents of self-sacrifice 58
Figure 3.
Self-sacrifice
100
100
90
80
67.9
70
60
50
Control Condition
48.5
Time Pressure
40
30
20
10
Letting 5
ingroup
members die
6.7
0
Weakly fused
Strongly fused
Note: Values for weakly vs. strongly fused participants were ± 1 SD from the mean (M =
2.47, SD = 1.22).
Antecedents of self-sacrifice 59
Figure 4.
Self-sacrifice
100
90
76.3
80
71.9
70
60
46.2
Control
50
35
40
Personal identity activation
Social identity activation
30
20
10.3
10
Letting 5
ingroup 0
members die
0
Weakly fused
Strongly fused
Note: Values for weakly vs. strongly fused participants were ± 1 SD from the mean (M =
2.31, SD = 1.22).
Antecedents of self-sacrifice 60
Figure 5
Self-sacrifice
100
90
79.7
80
76.1
70
57
60
50
Self-sacrifice for 1 ingroup
member
40
Self-sacrifice for 5 ingroup
members
32.1
30
20
10
Letting 1 or 5
0
ingroup
members die
Weakly fused
Strongly fused
Note: Values for weakly vs. strongly fused participants were ± 1 SD from the mean (M =
2.82, SD = 1.19).